Machine learning-assisted dual-mode intelligent biosensor for sarcosine detection based on fluorescent nanozymes

Area of Science:

• Biomedical Engineering
• Nanotechnology
• Analytical Chemistry

Background:

• Early prostate cancer detection is crucial but challenging.
• Sarcosine is an emerging biomarker for prostate cancer.
• Developing sensitive and cost-effective biosensors is a priority.

Purpose of the Study:

• To create a machine learning-assisted dual-mode biosensing platform.
• To detect sarcosine (Sar) ultrasensitively for prostate cancer screening.
• To integrate nanozyme catalysis with artificial intelligence for improved diagnostics.

Main Methods:

• Synthesized iron-doped carbon quantum dots (Fe_0.45-CDs) using microwave-assisted strategy.
• Developed a dual-mode sensor combining colorimetric and ratiometric fluorescent readouts.
• Employed machine learning algorithms (KNN) to analyze RGB features from smartphone images for sarcosine quantification.

Main Results:

• The Fe_0.45-CDs exhibited excellent fluorescence and peroxidase-like activity.
• The dual-mode platform achieved low limits of detection (0.922 μM colorimetric, 2.34 μM fluorometric).
• The KNN model demonstrated high prediction accuracy for sarcosine concentration in urine samples.

Conclusions:

• The developed Fe_0.45-CDs nanozyme biosensor offers a sensitive and reliable method for prostate cancer biomarker detection.
• The combination of nanozyme-bioenzyme cascade catalysis and AI enhances noninvasive cancer diagnosis.
• The platform shows promising clinical practicability due to its anti-interference capability and performance in real urine samples.